The legal brawl between Apple and OpenAI just got messier, and a lot more personal. OpenAI has published emails and texts that paint Apple’s lawsuit not as a principled defense of trade secrets, but as a sloppy, aggressive strike launched after a five-month silence. The core of OpenAI’s rebuttal is a cascade of Apple’s own admitted errors: it confused two Asian last names, emailed the wrong lawyer, and then an outside counsel falsely claimed he had spoken with OpenAI’s General Counsel on the phone.
Apple initially claimed it contacted OpenAI in February and got no response. The released email chain tells a different story. An attorney from Weil, Gotshal & Manges emailed the wrong person, Che Chang instead of the intended recipient, and then followed up by thanking Chang for a phone call that never happened. OpenAI’s General Counsel flagged the lie to Apple’s in-house team immediately. In a subsequent email, the Weil partner acknowledged the error and told OpenAI they were “resolving any issues.” After that note, OpenAI says it heard absolutely nothing until Apple filed suit in July. That’s a strange way to resolve issues.
OpenAI also takes direct aim at the accusations against two former Apple employees, Chang Liu and Tang Tan. The company says Apple now concedes that its own employees reached out to Liu after he left the company, asking for help locating files. OpenAI frames the “residual access” Apple complains about as a systemic failure on Apple’s part to properly cut off system access for departing staff—a problem that leaves ex-employees with files they didn’t want and didn’t know they had. Regarding Tang Tan, a 24-year Apple veteran, OpenAI states he has always been unequivocal internally: the team does not want, and must not use, any confidential information from other companies.
What’s striking here is the tone. OpenAI doesn’t hide its frustration, calling the lawsuit “careless, aggressive and oddly personal.” It says it offered to work with Apple to clear the air, but Apple instead chose to “change their narrative” and make vague accusations. The underlying message from OpenAI is blunt: we don’t have your secrets, we don’t want them, and this entire preliminary injunction is a theatrical move built on a foundation of factual errors. The documents are now public, and they make for an awkward read for a company that built its brand on sweating the details.
